Overview

Containers deliver a solution that sits somewhere between IaaS (GCE) and PaaS (App Engine).

GCE - VMs - An App runs on an OS running on a Hypervisor running on HW.

App Engine - An App runs on a service delivering Data, Cache, Storage, DBs, NW etc - very abstracted from the HW.

Containers - An App runs in a container which provides an OS delivered by the container engines which utilises the HW OS.

Kubernetes manages the execution of containers over a cluster of nodes (running on servers).

Cloud Build is a GCP container build tool.

Docker

Docker File: (pg_server)

FROM ubuntu:18.10
RUN apt-get update -y &&\
    apt-get install -y python-pip python3-dev
COPY requirements.txt /app/requirements.txt
WORKDIR /app
RUN pip3 install -r requirements.txt
COPY . /app
ENTRYPOINT ["python3", "app.py"]

> docker build -t pg_server .

> docker run -d pg_server

Kubernetes

> gcloud container clusters create k1

Pods

A Pod has a unique IP address.

> kubectl run nginx --image=nginx:1.15.7

> kubectl get pods

> kubectl expose deployment nginx --port=80 --type=LoadBalancer

A service groups pods together with a single point of access (a fixed IP).

> kubectl get services

Scaling is easy.

>kubectl scale nginx --replicas=3

Auto scale when CPU hits 80%.

kubectl autoscale nginx --min-10 --max=15 --cpu=80

Kubernetes adopts a Declarative Approach

Uses config files and not commands.

kubectl apply -f nginx-deployment.yaml

kubectl get replica

kubectl get pod

kubectl get deployments

kubectl get services

Anthos

Hybrid / Multi Cloud

Kubernetes with GKE on prem`.

Both can read containers from the GCP Marketplace.

This page was generated by GitHub Pages. Page last modified: 22/12/04 17:45